Photoinduced electron transfer reactions of 3H-pyrazole derivatives. Formation of solvent adduct by specific sensitizer

Article abstract of DOI:10.1246/bcsj.76.1227

3H-pyrazoles were photolyzed in the presence of 2,4,6-triphenylpyrylium tetrafluoroborate (TPP⁺) or 9,10-dicyanoanthracene (DCA) as a sensitizer. The obtained products were different depending on the sensitizers employed. The cyclopropene deriv

Full text of DOI:10.1246/bcsj.76.1227

ꢀ 2003 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 76, 1227–1231 (2003) 1227
Photoinduced Electron Transfer Reactions of 3H-Pyrazole Derivatives.
Formation of Solvent Adduct by Specific Sensitizer
ꢀ
ꢀ
Takashi Karatsu, Nobuo Shiochi, Takumi Aono, Nobukazu Miyagawa, and Akihide Kitamura
Department of Materials Technology, Faculty of Engineering, Chiba University, Yayoi-cho, Inage-ku, Chiba 263-8522
(Received October 17, 2002)
3H-pyrazoles were photolyzed in the presence of 2,4,6-triphenylpyrylium tetrafluoroborate (TPPþ) or 9,10-dicya-
noanthracene (DCA) as a sensitizer. The obtained products were different depending on the sensitizers employed. The
cyclopropene derivatives were produced in the cases of the DCA and TPPþ sensitizations and a solvent adduct was ob-
tained for the TPPþ sensitization reaction. The difference in products is explained by the rates of the back electron trans-
fer from the electron accepted sensitizers to the radical cation intermediates. In the DCA sensitization, the absorption
band of the diazoalkene was observed but in the TPPþ sensitization it was not observed. From these results, it is ap-
parent that the back electron transfer rates of the DCA sensitization reactions are faster than those of the TPPþ sensi-
tization reactions.
Many photoinduced electron transfer (PET) reactions of
cyclic azoalkanes have been studied, however, in most cases,
the photodenitrogenation products were obtained in the PET
reaction similar to the direct photolysis.1 Few characteristic
products are hardly obtained by the PET reaction. In the
PET reaction of a five-membered cyclic azoalkane, pyrazoline,
the rearranged product was produced with the cyclopropane
which is the only product in the thermolysis and in the direct
photolysis.2 This difference in products was explained by the
stability of the radical cation intermediates. In our previous
short communication, we reported the PET reaction of 3H-pyr-
azoles which have an unsaturated carbon–carbon double bond
with an azo chromophore in its five-membered skeleton.3 The
direct photolysis of the 3H-pyrazoles had been already report-
ed and the corresponding cyclopropene derivatives were pro-
duced after the observation of the diazoalkene precursors.4
In the paper, we showed the interesting result that the use of
different sensitizers induced different products. In the 9,10-di-
cyanoanthracene (DCA) sensitization, only the cyclopropene
was produced, but in the 2,4,6-triphenylpyrylium tetrafluoro-
borate (TPPþ) sensitization, an adduct with the solvent, aceto-
nitrile, was produced with the cyclopropene. In the PET reac-
tions, the formations of solvent adducts have been reported by
several groups.5 Such formation means that the lifetime of the
intermediate radical cation is long enough to react with the sol-
vent.
Table 1. Redox Potentials of 3H-Pyrazoles, Their Quench-
ing Rate Constants (kq) by TPPþ and DCA, and Their Free
Energy Change (ꢀG) Relationship
Eox/V Ered/V ꢀG/kcal molꢂ1
kq/L molꢂ1 sꢂ1
TPPþ
DCA
Compound
vs SCE
TPPþ DCA
1
2
3
2.05 ꢂ1:45 ꢂ11:8 +1.2 1:6 ꢃ 1010 6:4 ꢃ 109
2.05 ꢂ1:59 ꢂ11:8 +1.2 1:1 ꢃ 1010 1:3 ꢃ 109
2.02
—
ꢂ12:5 +0.5 1:7 ꢃ 1010 1:5 ꢃ 109
by 3H-pyrazoles, but the quenching rate constants of DCA
by the 3H-pyrazoles were five to ten times lower than those
of TPPþ. The differences in the free energy (ꢀG) calculated
from the measured oxidation potentials of 1 and 2, using the
Rehm–Weller equation,6 are shown in Table 1. The difference
in kqs can be explained from these values. For TPPþ, ꢀG val-
ues are exergonic and an electron certainly transfers from a
molecule of the 3H-pyrazoles to the excited singlet TPPþ mo-
lecule at the diffusion-controlled rate constants. On the other
hand, for DCA, the ꢀG values are slightly endergonic and the
kq values must reflect these relations. In Table 1, the reduction
potentials of 1 and 2 are also listed. These values are not used
to calculate ꢀG, but they are useful for understanding the re-
dox correlation between the sensitizers and 3H-pyrazoles.
Furthermore, the oxidation potential of 3 and the quenching
rate constants by 3 are listed in Table 1.
In this paper, we show the total reaction mechanism of the
3H-pyrazoles which is confirmed from some experiments in-
volving both the DCA and the TPPþ sensitization.
The electron transfer between the sensitizer and 3H-pyra-
zole is also supported by the observation of a pyranyl radical
(TPPꢁ) during the laser flash photolysis experiment. The tran-
sient absorption spectrum was observed at around 550 nm after
the laser pulse and it lived for ca. 50 ms; this is assigned to the
pyranyl radical absorption.7;8 The observation of the pyranyl
radical means that an electron transferred from a ground state
3H-pyrazole to an excited singlet state TPPþ. However, the
electron donated species, the 3H-pyrazole radical cation or
its derived species, was not observed as the transient absorp-
Results and Discussion
In order to investigate the interaction between the excited
singlet sensitizers and 3H-pyrazoles, fluorescence quenching
experiments were performed. The obtained quenching rate
constants (kq) are summarized in Table 1. The fluorescence
of TPPþ was quenched at the nearly diffusion rate constants

1228 Bull. Chem. Soc. Jpn., 76, No. 6 (2003)
PET Reactions of 3H-Pyrazoles
Fig. 2. Decomposition of 1 and formation of products in
the presence of TPPþ by the irradiation of 405-nm light.
1 (— —); 3 (— —); 4 (— —).
was < 50 ns. This very short lifetime must be caused by the
fast BET under the contact ion pairs. On the other hand, for
TPPþ, the BET does not efficiently occur because of producing
the solvent separated pairs more rapidly.10 The formation of 3
is induced by the decomposition of 5 and 6. These reaction
mechanisms are not certain, but 5 may be decomposed by
DCA sensitization because 5 cannot directly absorb the light
owing to the overlap with the absorption of DCA. On the other
hand, the sensitized and/or direct photolysis of 6 is presumed
to occur by the absorption because the absorption band of 6
is isolated from that of DCA.
Fig. 1. Absorption spectra of 1 (4:4 ꢃ 10ꢂ3 mol dmꢂ3) in
the presence of DCA (4:2 ꢃ 10ꢂ4 mol dmꢂ3) in acetoni-
trile for various irradiation times. Inset: Expansion of
the wavelength between 400 and 640 nm.
During the TPPþ sensitization reactions, both 1 and 2
afforded the same products, 3 and 4. The characteristic com-
pound 4 would be an adduct of the 1,3-radical cation, 7, with
the solvent, acetonitrile. The production of the same com-
pound from the different reactions means the presence of the
same intermediate or precursor. In order to resolve the reac-
tion mechanism, we examined the irradiation time correlations
of these product yields. In the TPPþ sensitization reaction of
1, during the primary stage, the cyclopropene 3 was generated
according to the consumption of the starting material. After 1
h, the yield of 3 reached a plateau, then the yield of 4 gradually
increased after a short induction time (Fig. 2). For the reaction
of 2, 4 was generated rather than 3 at the initial stage (Fig. 3).
The controlled experiments showed that the TPPþ sensitiza-
tion reaction of 3 gave 4, however, only the consumption of
4 and no formation of 3 were observed by the TPPþ sensitiza-
tion of 4.
From the results of these experiments, we confirmed the to-
tal reaction mechanism of the DCA and TPPþ sensitization of
the 3H-pyrazoles, 1 and 2, as shown in Scheme 1. During the
DCA sensitization reaction of 1, the electron transfer occurred
to generate the radical cation of 1 and produce the radical cat-
ion 8 by the carbon–nitrogen bond cleavage. Then, the forma-
tion of the diazoalkene 5 is observed by the fast BET process
between 8 and the DCA radical anion, and next nitrogen 3 is
eliminated. No further reaction occurred in the case of
DCA. On the other hand, for the TPPþ sensitization reaction
of 1, the electron transfer occurred to generate the radical cat-
ion of 1, and produce the radical cation 8 by the carbon–nitro-
tions probably owing to an overlap with the absorption of
TPPþ.
In the DCA sensitizations, 1 and 2 were efficiently decom-
posed and produced the same compound, 3. The yields were
45% and 85%, respectively. During the direct photolyses of
1 and 2, 3 was obtained and the presence of precursors, di-
azoalkene derivatives 5 and 6, were confirmed from the UV
absorption spectra which had maximum wavelengths of 420
nm and 510 nm, respectively.4 These diazoalkene compounds
were generated by the radical cleavages of the carbon–nitrogen
bonds of 1 and 2. During the TPPþ sensitization reactions, we
did not observe the diazoalkene precursors, but surprisingly we
observed the generation of 6 for the DCA sensitization of 2.
The absorption band, which had a maximum at 510 nm, was
gradually increased and then decreased by the irradiation as
shown in Fig. 1. We believe 5 must be also generated during
the DCA sensitization of 1, although it could not be observed
in the absorption spectra due to the overlapping with the ab-
sorption band of DCA. The mechanism of these reactions
looks to be the same as the direct photolysis and the thermoly-
sis, which means that the back electron transfer (BET), from
the DCA anion radical to the C–N bond cleaved cation radical,
8, 9 or 10, occurred under the contact ion pairs.7 The laser
flash photolysis experiments supported this estimation. The
transient absorption spectrum of the radical anion of DCA9
was observed after the laser pulse in the presence of DCA
and biphenyl, and its lifetime was 8 ms. However, in the pre-
sence of DCA and 1, the lifetime of the anion radical of DCA

T. Karatsu et al.
Bull. Chem. Soc. Jpn., 76, No. 6 (2003) 1229
Scheme 1.
than that of 1. There are two reaction paths. One is the path
via the intermediate 9 and the other is the path via the inter-
mediate 10. The difference of the cleaved carbon–nitrogen
bonds of the radical cation of 2 induces two kinds of radical
cations, 9 and 10. After the generation of 9, the reaction pro-
ceeds similarly to the case of 8. In the DCA sensitization re-
action, the generation of 6 is observed and the product is the
cyclopropene 3, and in the TPPþ sensitization reaction the cy-
clopropene 3 and 2H-pyrrole 4 are obtained. The fate of 10 is
different from those of 8 and 9. In the DCA sensitization re-
action, the biradical 11 is produced by the BET between 10
and the DCA radical anion, and the cyclopropene 3 and the
starting material, 3H-pyrazole 2 would be produced. In the
TPPþ sensitization reaction, the radical cation 10 eliminates
nitrogen and the 1,3-radical cation 7 is directly produced,
and then the cyclopropene 3 and 2H-pyrrole 4 are obtained.
All the results can be clearly explained by the reaction
mechanism of Scheme 1 except for one problem, why the
mechanism was different in the reactions of 1 and 2. In
Scheme 1, it is shown that the decomposition of 1 proceeds
through the intermediate of the bond cleaved radical cation,
8. On the other hand, during the reaction of 2, there are two
patterns of carbon–nitrogen bond cleavage and radical cations
9 and 10 are generated. This difference in the bond cleavage
pattern is supported by the AM1 calculation. In the model cal-
culations for the cationic reactions in vacuum, therefore, the
activation energies seem to be too high for efficient
Fig. 3. Decomposition of 2 and formation of products in
the presence of TPPþ by the irradiation of 405-nm light.
2 (— —); 3 (— —); 4 (— —).
gen bond cleavage. Then, 8 eliminates nitrogen and the cyclo-
propene 3 is produced by the BET and the recyclization
(CYCL) of the generated 1,3-radical cation. Furthermore, in
the presence of TPPþ, the carbon–carbon bond of 3 is cleaved
by the PET reaction to form the isomeric 1,3-radical cation 7,
then, acetonitrile, which is a weak nucleophile, adds to 7, re-
sulting in the formation of 2H-pyrrole 4.
The sensitized reaction mechanism of 2 is more complex

1230 Bull. Chem. Soc. Jpn., 76, No. 6 (2003)
PET Reactions of 3H-Pyrazoles
Fig. 4. Potential energy diagrams of radical cations derived
from 1 by AM1 calculation. The values of the closed cir-
cles are calculated, but the value of the open circle was not
obtained, because another 1,3-radical cation is more stable
and the result of the calculation corresponds to the more
stable 1,3-radical cation with the eliminated nitrogen.
Fig. 5. Potential energy diagrams of radical cations derived
from 2 by AM1 calculation. The values of the closed cir-
cles are calculated, but the value of the open circle was not
obtained, because another 1,3-radical cation is more stable
and the result of the calculation corresponds to the more
stable 1,3-radical cation with the eliminated nitrogen.
reactions. These values for cationic reactions must take much
smaller values in polar solvents. The bond cleavage of the ra-
dical cation of 1 is endothermic and the activation energy to
generate the radical cation of 8 from 1 is 23 kcal molꢂ1, but
that of the opposite side C–N bond cleavage requires about
40 kcal molꢂ1 (Fig. 4). This is the reason why the reaction
proceeds through the intermediate 8, and then the 1,3-radical
cation is generated by the elimination of nitrogen. For the re-
action of 2, the activation energies to generate the radical cat-
ion 9 or 10 are quite similar, 26 and 27 kcal molꢂ1, respective-
ly (Fig. 5). Therefore, the reaction proceeds through two
paths, one is the path via the radical cation 9 and the other
is the path via the radical cation 10.
More important information from the calculation is obtained
for the propenyl and cyclopropenyl radical cations. If the cy-
clopropenyl radical cation of 3 is generated by the PET reac-
tion of 3, the C–C (benzyl carbon) bond is easier to cleave than
the other C–C bonds (Fig. 6). This lower activation energy
leads to the formation of the acetonitrile adduct 4.
Fig. 6. Potential energy diagrams of radical cations derived
from 3 by AMI calculation.
Synthesis of Methyl 2,2-Dimethyl-4-phenyl-2H-pyrole-3-
carboxylate (4). A solution of 1 mmol of 1 in 25 mL of aceto-
nitrile with 30 mg of TPPþ was distributed into 6 irradiation tubes,
and then degassed by freeze-thaw cycles. The degassed samples
were irradiated at 405 nm for 10 h, and then the samples were col-
lected and concentrated. After removal of TPPþ through a short
column chromatography (AcOEt), the residue was concentrated
and purified by preparative TLC (CHCl3/AcOEt). A colorless
oil was obtained in 15% yield. 1H NMR (270 MHz, CDCl3) ꢀ
7.42–7.38 (m, 2H, Ar), 7.25–7.19 (m, 3H, Ar), 3.61 (s, 3H,
CO2Me), 2.12 (s, 3H, MeC=N), 1.45 (s, 6H, Me). NOE was ob-
served between the aromatic proton and the methyl proton derived
from the MeCN unit. HRMS (EI) Calcd for C15H17NO2
243.1259, found 243.1251.
We are interested in the PET reactions of cyclic azo com-
pounds, and especially the behavior of the generated radical
cations. In this case we found a fairly stable 1,3-radical cation
which can react with the solvent, acetonitrile. We are now
studying more stable radical cations which are generated from
cyclic azo compounds.
Experimental
Materials. The sensitizers were purified by recrystallizaiton,
and dehydrated acetonitrile (Kanto Chemical Co.) was used.
Methyl 3,3-dimethyl-4-phenyl-3H-pyrazole-5-carboxylate (1)
and methyl 3,3-dimethyl-5-phenyl-3H-pyrazole-4-carboxylate (2)
were prepared by the cycloaddition of 2-diazopropane to methyl
phenylpropiolate.4 Methyl 3,3-dimethyl-1-phenylcyclopropene-
2-carboxylate (3)4 was isolated from the pyrolyzed sample of 1.
Fluorescence Quenching Experiments. The electron transfer
rate constant was determined by the change in the fluorescence in-
tensity of DCA (1:9 ꢃ 10ꢂ4 mol dmꢂ3) or TPPþ (5:2 ꢃ 10ꢂ4
mol dmꢂ3) during the addition of various concentrations of the
3H-pyrazoles (0{1 ꢃ 10ꢂ2 mol dmꢂ3
) in acetonitrile under
degassed condition.

T. Karatsu et al.
Bull. Chem. Soc. Jpn., 76, No. 6 (2003) 1231
Redox Potential Measurements. The oxidation and the re-
duction potentials of the 3H-pyrazoles were measured by cyclic
voltammetry (BAS CV-1B) at a platinum electrode in argon-satu-
rated dry acetonitrile with 0.1 mol dmꢂ1 tetrabutylammonium tet-
rafluoroborate as the supporting electrolyte. The scan speed was
200 mV sꢂ1 and the reference electrode was a SCE. The obtained
peaks were irreversible, and the potentials were determined from
the half-peak potentials.
Transient Absorption Measurements. 2:3 ꢃ 10ꢂ5 mol dmꢂ3
of TPPþ and 2:2 ꢃ 10ꢂ3 mol dmꢂ3 of 1, 2:7 ꢃ 10ꢂ3 mol dmꢂ3 of
DCA and 2:0 ꢃ 10ꢂ1 mol dmꢂ3 of biphenyl, 2:7 ꢃ 10ꢂ3 mol dmꢂ3
of DCA and 1:7 ꢃ 10ꢂ3 mol dmꢂ3 of 1, respectively, are dissolved
in acetonitrile; argon gas was bubbled for 20 min and the transient
absorptions were measured. They were measured by excitation at
355 nm (Continuum SL I-10, 6 ns fwhm, 20 mJ per pulse) with a
detection system (Tokyo Instruments) composed of a multichan-
nel diode array (Princeton IRY-512G: 18 ns gate) with a SPEX
270 M monochromometer (resolution: 0.3 nm/channel).11
A part of this work was supported by a Grant-in-Aid for Scien-
tific Research from the Ministry of Education, Culture, Sports,
Science and Technology and the Futaba Denshi Foundation.
References
1
a) W. Adam, U. Denninger, R. Fnzel, F. Kita, H. Platsch,
H. Walter, and G. Zang, J. Am. Chem. Soc., 114, 5027 (1992).
b) W. Adam and J. Sendelbach, J. Org. Chem., 58, 5316 (1993).
c) H. Ikeda, T. Minegishi, and T. Miyashi, J. Chem. Soc., Chem.
Commun., 1994, 297. d) H. Ikeda, T. Minegishi, Y. Takahashi,
and T. Miyashi, Teterahedron Lett., 37, 4377 (1996).
2
a) T. Karatsu, H. Hotta, and A. Kitamura, J. Chem. Soc.,
Chem. Commun., 1991, 1451. b) T. Karatsu, H. Itoh, T. Kikunaga,
Y. Ebashi, H. Hotta, and A. Kitamura, J. Org. Chem., 60, 8270
(1995).
3 N. Miyagawa, T. Karatsu, and A. Kitamura, Chem. Lett.,
1997, 1005.
Reactions. Direct Photolyses of 1 and 2:
2:9 ꢃ 10ꢂ3
4 M. Frank-Neumann and C. Buchecker, Tetrahedron Lett.,
1, 15 (1969).
mol dmꢂ3 of 1 or 2 was dissolved in acetonitrile. After being de-
gassed by the freeze-thaw in a tube with a UV cell, the sample was
irradiated by 366-nm light using a 1-kW high pressure mercury
lamp through a glass filter. The reactions were pursued by UV
and GC analysis.
5
a) T. A. Zona and J. L. Goodman, Tetrahedron Lett., 33,
6093 (1992). b) K. Mizuno, K. Nire, H. Sugita, and Y. Otsuji, Tet-
rahedron Lett., 34, 6563 (1993). c) H. J. P. deLijser and D. R.
Arnold, J. Org. Chem., 62, 8432 (1997).
DCA Sensitized Photolyses of 1 and 2: 1 or 2 and DCA were
dissolved in acetonitrile to make a solution having a concentration
of 4:4 ꢃ 10ꢂ3 mol dmꢂ3 of 1 or 2 and 4:2 ꢃ 10ꢂ4 mol dmꢂ3 of
DCA which was nearly saturated. After being degassed by the
freeze-thaw, the sample was irradiated by 405-nm light using a
1-kW high pressure mercury lamp through a CuSO4 + NaNO2
+ aqueous ammonia solution filter.12 The reactions were moni-
tored by GC analysis.
TPPþ Sensitized Photolyses of 1 and 2: 1 or 2 and TPPþ
were dissolved in acetonitrile to make a solution having a concen-
tration of 5:7 ꢃ 10ꢂ3 mol dmꢂ3 of 1 or 2 and 1:4 ꢃ 10ꢂ3
mol dmꢂ3 of TPPþ. After being degassed by the freeze-thaw,
the sample was irradiated by 405-nm light using a 1-kW high
pressure mercury lamp through the solution filter. The reaction
was monitored by GC analysis.
6 a) D. Rehm and A. Weller, Ber. Bunsen-Ges. Phys. Chem.,
73, 834 (1969). b) D. Rehm and A. Weller, Isr. J. Chem., 8, 259
(1970).
7 T. Karatsu, H. Kobayashi, E. Shinkai, and A. Kitamura,
Chem. Lett., 1992, 2131.
8 N. Miyagawa, T. Karatsu, Y. Futami, T. Kunihiro, A.
Kiyata, and A. Kitamura, Bull. Chem. Soc. Jpn., 69, 3325 (1996).
9 I. R. Gould, D. Ege, J. E. Moser, and S. Farid, J. Am.
Chem. Soc., 112, 4290 (1990).
10 T. Karatsu, K. Kanayama, M. Takahashi, N. Ishigoka, K.
Fukui, and A. Kitamura, Heteroatom Chem., 12, 269 (2001).
11 T. Karatsu, H. Itoh, N. Yoshikawa, A. Kitamura, and K.
Tokumaru, Bull. Chem. Soc. Jpn., 72, 1839 (1999).
12 J. Eriksen and C. S. Foote, J. Am. Chem. Soc., 102, 6083
(1980).